The Sanctions ParadoxEconomic Statecraft andInternational Relations

Daniel W. Drezner

published by the press syndicate of the university of cambridgeThe Pitt Building, Trumpington Street, Cambridge, United Kingdom

cambridge university pressThe Edinburgh Building, Cambridge, CB2 2RU, UK

http://www.cup.cam.ac.uk40 West 20th Street, New York, NY 10011±4211, USA

http://www.cup.org10 Stamford Road, Oakleigh, Melbourne 3166, Australia

# Daniel W. Drezner 1999

This book is in copyright. Subject to statutory exception and to the provisionsof relevant collective licensing agreements, no reproduction of any part maytake place without the written permission of Cambridge University Press.

First published 1999

Printed in the United Kingdom at the University Press, Cambridge

Typeset in Palatino 10/12.5pt [ce ]

A catalogue record for this book is available from the British Library

Library of Congress cataloguing in publication data

Drezner, Daniel W.The sanctions paradox: economic statecraft and internationalrelations / Daniel W. Drezner.

p. cm. ± (Cambridge studies in international relations: 65)ISBN 0 531 64332 5 (hardback) ± ISBN 0 521 64415 1 (paperback)1. Economic sanctions. 2. International economic relations.3. Economic sanctions ± Case studies. I. Title. II. Series.HF1413.5.D74 1999 327.1'17±dc21 98±39105 CIP

ISBN 0 521 64332 5 hardbackISBN 0 521 64415 1 paperback

Contents

List of ®gures page xiiList of tables xiiiAcknowledgments xv

1 Introduction 1A tale of two cases 1The argument 4Why economic coercion matters 6The literature 9Flaws in the literature 18The methodology 21The rest of the book 22

Part I Theory and data 25

2 A model of economic coercion 27The assumptions 28The model 35A caveat: switches in preferences 47Another caveat: non-negotiable demands 48Choosing between carrots and sticks 50Conclusions and implications 53Appendix: proofs of lemmas 55

3 Plausibility probes 59First impressions 60Statistical studies of sanctions initiation 62

ix

Statistical studies of sanctions outcomes 68Case study 1: the grain embargo 74Case study 2: the pipeline sanctions 80Case study 3: economic sanctions and human rights 88Summary 99

4 Statistical tests 102Selecting the appropriate sample 102Operationalizing the variables 106Testing the causes of sanctions initiation 114Testing the causes of sanctions outcomes 121Summary 127

Part II Economic coercion in the former Soviet Union 129

5 Russian power and preferences 131The motivation 131Russian policy preferences 134Russia's ability to coerce 140Con¯ict expectations with Russia 147Predicting the outcomes 150

6 The extent of NIS concessions 153Introduction 153Belarus 154Kazakhstan 160Turkmenistan 169Kyrgyzstan 173Armenia 176Tajikistan 179Uzbekistan 183Moldova 187Georgia 193Ukraine 198Azerbaijan 208Lithuania 215Latvia 219Estonia 224Conclusion 229

Contents

x

7 Evaluating the evidence 231Introduction 231Coding the data 232Predictions and outcomes 233A Boolean analysis 239Some ®nal observations 245

Part III Choosing between carrots and sticks 249

8 Economic statecraft and nuclear proliferation on theKorean Peninsula 251Carrots and sticks 251The United States and South Korea's nuclear program 254The United States and North Korea's nuclear program 275Implications 303

Part IV Conclusion 305

9 Conclusions, implications, speculations 307A review 307Additional insights 309A critique 311Policy implications 312Questions for future research 317

References 322Index 336

Contents

xi

Figures

2.1 The coercion game page 372.2 The sender's coercion decision 422.3 The effect of con¯ict expectations on concession size 462.4 The statecraft game 505.1 Russia and the NIS 132

xii

Tables

1.1 The predicted pattern of economic coercion 51.2 Existing approaches to economic coercion 142.1 The predicted pattern of economic coercion 544.1 Frequency distribution of concession size 1094.2 Dependent and explanatory variables used in

statistical tests 1124.3 Control variables used in statistical tests 1134.4 Distribution of the gap in opportunity costs 1164.5 Testing the coercion condition 1174.6 An alternative test of the coercion condition 1184.7 Test of the domestic politics hypothesis of sanctions

initiation 1204.8 Test of the signaling hypothesis 1214.9 Testing the concession function 123

4.10 Alternative test of the determinants of concession size 1254.11 The determinants of sanctions length 1264.12 Alternative test of the determinants of sanctions length 127

5.1 An introduction to the NIS 1335.2 Distribution of land-based nuclear weapons in the

former Soviet Union 1375.3 Strategic military bases in the former Soviet Union 1385.4 NIS trade dependence on Russia, 1994 1435.5 NIS debt to Russia, June 1992 1445.6 Energy production in the former Soviet Union 1465.7 NIS energy dependence on Russia 1465.8 NIS vulnerability to Russian economic coercion 1475.9 Predicting the pattern of economic coercion in the NIS 1517.1 Coding NIS concessions 232

xiii

7.2 Summary of coercion episodes in the formerSoviet Union 234

7.3 The pattern of Russia's coercion attempts 2377.4 The pattern of signi®cant concessions from the NIS 2377.5 The pattern of signi®cant and moderate concessions

from the NIS 2387.6 The statistical determinants of NIS concessions 2397.7 Boolean truth table for signi®cant concessions 2417.8 Boolean truth table for signi®cant and moderate

concessions 2438.1 ROK energy imports, 1960±75 2618.2 ROK electric-generation capacity, 1956±2000 2618.3 Contributions to ROK military expenditures, 1969±79 2748.4 US military assistance to ROK, 1971±9 2758.5 North Korea's economic health, 1990±4 277

xiv

List of tables

1 Introduction

A tale of two cases

On August 6, 1990, the United Nations Security Council voted toimpose multilateral economic sanctions against Iraq in response toSaddam Hussein's invasion of Kuwait. By any conventionalmeasure, these sanctions achieved the greatest degree of inter-national cooperation in modern history. Iraq suffered damages equalto roughly half its pre-war gross national product, a cost faroutweighing any other sanctions attempt in this century. Despite thesevere economic dislocations suffered by the Iraqi regime, it refusedto surrender any Kuwaiti territory. Ignoring claims that the sanc-tions would have worked with time, the US-led coalition decided toretake Kuwait through military force. On April 3, 1991, the SecurityCouncil voted to extend the sanctions regime until Iraq compliedwith additional demands to reveal its weapons of mass destructionprogram, recognize the border with Kuwait, and pay reparations.An unstated but desired demand was the removal of SaddamHussein from power. Iraq has labored under the UN sanctionsregime for eight years. As a result, infant mortality rates haveincreased sevenfold, annual in¯ation rose to over 4,000 percent, andper capita income has fallen to less than half pre-war levels. In theface of continued economic losses and bellicose US rhetoric, theIraqi regime has only acquiesced to UN demands when additionalmilitary threats have been made. On every issue area, when theonly pressure is economic sanctions, Iraq has not budged. Domes-tically, Saddam's regime shows no signs of falling; if anything, thesanctions regime has strengthened it. Despite the most potent

1

sanctions in history, economic coercion has failed to produce anysigni®cant Iraqi concessions.1

In late August 1991, the United States was trying to cajole all therelevant players in the Middle East into a multilateral peace confer-ence in Madrid. A blocking point was the Israeli construction of newhousing in the occupied territories of the West Bank and Gaza. TheIsraeli government of Yitzhak Shamir sought to acquire US-ensured®nancing to cover additional expenses. The Bush administrationresponded by delaying and then refusing to grant $10 billion in loanguarantees until the Israelis froze the building of all new settlementsin the territories. Shamir agreed to the delay before the Madrid peaceconference, but refused to concede on the issue of new housing. USSecretary of State James Baker told the Israeli government that theloan guarantees were conditional on the freeze in housing construc-tion in the occupied territories. With Shamir reluctant to yield,Washington used economic pressure to force a change in government.US and Israeli authorities agree that Shamir's refusal to concede costhim and his Likud party the June 1992 elections. In August of 1992,the new Labor government, led by Yitzhak Rabin, acquiesced to theUS terms and agreed to freeze the building of all new settlements inreturn for the loan guarantees. Shamir's Defense Minister and cam-paign manager later described the episode as an unprecedentedexample of US interference in Israeli domestic politics.2 Nevertheless,the Israeli government acquiesced to US economic pressure, pavingthe way for the historic 1993 accord between Israel and the PalestineLiberation Organization.

Although the Iraqi case has been the focus of more attention, bothepisodes are examples of economic coercion. I de®ne economic coer-cion as the threat or act by a nation-state or coalition of nation-states,called the sender, to disrupt economic exchange with another nation-state, called the target, unless the targeted country acquiesces to anarticulated political demand. The disrupted exchange could include

1 See Hufbauer, Schott, and Elliott (1990a), pp. 283±98 on the pre-war UN sanctions. Onthe postwar sanctions regime and its effect on Iraq's population, see Reuther (1995) andthe Economist, ``Iraq: more medicine please,'' March 7, 1998. On how the sanctions havestrengthened Hussein's grip on power, see ``CIA says Saddam Hussein emergedstronger,'' Boston Globe, September 20, 1996, p. A2, and Robert Wright and JohnDaniszewski, ``Hussein may be main bene®ciary of UN aid effort,'' Los Angeles Times,March 4, 1998, p. A1.2 See Arens (1995), pp. 1, 301. See Baker (1995), pp. 540±57, for the US take on theepisode.

2

Introduction

trade sanctions, boycotts, aid suspensions, freezing of ®nancial assets,or the manipulation of tariff rates.3

The two cases suggest several puzzles that need to be addressed.First, why did unilateral US pressure on Israel succeed where multi-lateral UN pressure on Iraq failed? The existing explanations abouteconomic coercion are of little use. Experts predicted that the sanc-tions against Iraq would be successful.4 The case had all the attributestraditionally associated with successful sanctions: the target incurredsigni®cant economic damage; the sanctions effort was backed by aninternational institution; all major trading partners cooperated withthe UN resolutions; and the threat of military force behind thesanctions effort was clearly signaled. Yet the coalition leaders believedthat economic statecraft would not be enough to extract the desiredconcessions; sanctions failed to achieve their goals without theadditional use of force. By contrast, the Israeli case involved unilateralsanctions that imposed smaller though still signi®cant costs on thetarget and no threat of military force. Unlike the Iraqi case, Congresswas reluctant to threaten economic coercion, and in taking actionPresident Bush incurred the wrath of the American±Israeli PublicAffairs Committee, one of the most powerful foreign policy lobbies inthe United States. Despite all this, coercion was still successful. Whataccounts for the extent of the target country's concessions?

A second, less obvious, question is what prompts the initial attemptat economic coercion? In both cases, the target governments had takenactions that con¯icted with the policy preferences of the United States.While that is clearly part of the answer, it fails to explain why theUnited States chose to use economic coercion as its preferred policyoption. It could have done nothing; it could have tried traditionaldiplomacy; it could have offered economic inducements; it could haveresponded with immediate military intervention. Indeed, in the Iraqicase, the United States eventually showed a preference to use force

3 I will use the terms economic coercion, economic statecraft, and economic sanctionsinterchangeably in the interest of style, but they are technically different. Economicstatecraft includes the use of inducements as well as sanctions. In the public perception,economic sanctions are associated with trade-disrupting measures. The de®nition ofeconomic coercion includes cases of economic sanctions such as those against Iraq, but italso includes episodes such as the Israeli one, which is not commonly thought of as asanctions case.4 Gary Hufbauer and Jeffrey Schott testi®ed before Congress in December 1990 thatsanctions would compel Iraq to withdraw from Kuwait, although it could take morethan a year. See Pape (1997), n. 3.

Introduction

3

instead of relying on economic coercion. In the Israeli case, it couldhave offered a carrot, as it had done in the past. How do senderschoose among their policy options in an international crisis? Underwhat conditions will a sender attempt economic coercion?

The argument

This book is about the role of economic coercion in internationalrelations. It creates a model to explain the behavior of senders andtargets by taking into account their opportunity costs of deadlock andexpectation of future con¯ict with each other. Coercion alters theallocation of bene®ts by imposing costs on both the sender and targetcountries. The short-run costs of sanctions imposition are important tothe target and sender, but they are not the only factor. Conceding inthe face of economic coercion implies a redistribution of politicalassets between the target and sender. Nation-states care about thisredistribution if they think it will harm their bargaining position infuture con¯icts. This expectation of future con¯ict is translated into ashort-run concern for relative gains and reputation that varies withthe expectation of future threats or con¯icts in the bilateral relation-ship between the sender and target.

The expectation of future con¯ict has a contradictory effect oneconomic coercion. On the one hand, it makes senders more willing tothreaten economic sanctions. The greater the concern for relative gainsand reputation, the more likely the sender will prefer a ``stalemate'' or``deadlock'' outcome of disrupted economic exchange and attempt tocoerce. Ceteris paribus, senders will be eager to coerce adversaries, andreluctant to coerce allies.

The sender's enthusiasm does not translate into greater concessions.The second effect of con¯ict expectations is paradoxical and sur-prising. While a robust anticipation of future disputes might make thesender prefer a coercive strategy, it also reduces its ability to obtainconcessions. The target's con¯ict expectations determine the magni-tude of concessions. Facing an adversarial sender, the target will beworried about the long-run implications of acquiescing. Because itexpects frequent con¯icts, the target will be concerned about anyconcessions in the present undercutting its bargaining position infuture interactions. The sender might exploit the material or reputationeffects from these concessions in later con¯icts. When relative-gainsconcern is prominent, a concession represents a gain for the coercer

4

Introduction

and a loss for the coerced. When reputation is important, acquiescencebolsters the sender's credibility as a tough negotiator while weakeningthe target's reputation. With allies, this concern is less prominent,because the target anticipates fewer zero-sum con¯icts. Ceteris paribus,targets will concede more to allies than adversaries. Ironically, asender will obtain the most favorable distribution of payoffs when itcares the least about the relative distribution of gains.

Table 1.1 summarizes the predictions of a con¯ict expectationsmodel. Between adversaries, senders will be more willing to sanction,even if a target's costs of deadlock are only slightly greater than thesender's own costs. Despite these preferences, it will not be able toextract signi®cant concessions from the coercion attempt. Because thetarget is also concerned with the future implications of backing down,any concession is a double blow; not only does it lose in the short run,it grants the sender greater leverage in future disputes. While thepresence of con¯ict expectations might make the sanctioner preferdeadlock, it also makes acquiescence less palatable to the sanctioned.Thus, between adversaries, I expect to see sanctions that are fre-quently costly to the sender and produce only marginal concessions.

Between allies, the sender will be unwilling to threaten economiccoercion unless the gap in the costs of sanctions imposition is large.The target's costs of deadlock must be signi®cant and the sender'sown costs must be small for the sender to prefer a coercion attempt.Once this threshold is met, however, economic sanctions will befruitful. Because the target does not anticipate many future con¯icts, itwill care less about the material and reputational implications ofconceding, and more about the immediate costs and bene®ts. It will

Introduction

5

Table 1.1. The predicted pattern of economic coercion

Minimal Heightenedcon¯ict expectations con¯ict expectations

Large gap Signi®cant Moderatein costs concessions concessions

Small gap No coercion Minorin costs attempt concessions

concede more to avoid the costs of deadlock. Thus, between allies, Iexpect to see sanctions that are less costly to the sender, with moreproductive results.

A con¯ict expectations model can explain the dynamics of economiccoercion with more accuracy and parsimony than any existing expla-nation. A problem with the existing literature is that it has focused toomuch on the most publicized cases of economic sanctions, whichusually involve adversaries. Because of this sample bias, alternativeexplanations overlook less contentious but more successful coercionattempts between allies. In focusing on a limited subset of coercioncases, these writings have painted a distorted picture of economicsanctions. These arguments are not necessarily wrong, but their effectsare much smaller than their proponents claim. In contrast to a con¯ictexpectations model, they explain fewer cases, and less of the variationin outcomes.

Why economic coercion matters

Why should anyone care about economic statecraft? Two reasons, onefor the pragmatist and one for the theorist. The practical reason is thatthe incidence of economic sanctions has multiplied since the end ofthe cold war, without a similar increase in policy analysis. The esotericreason is that an examination of economic statecraft can illuminate thenature of power in international relations.

The use of economic statecraft in international relations has a longpedigree. The Athenian boycott of Megara helped to trigger thePeloponnesian war. The trading empires of Venice, Portugal, and theNetherlands used economic warfare to limit the power of their rivals.Early Anglo-American relations were consumed with issues ofeconomic diplomacy; a chief complaint in the Declaration of Indepen-dence is the ``Cutting off our Trade with all Parts of the World.''Woodrow Wilson believed that the ``economic, peaceful, silent deadlyremedy'' of economic sanctions could be used by the League ofNations to police international society. Nazi Germany was particularlyaggressive at cultivating economic dependency from its eastern Euro-pean neighbors. The US embargo of Japan in the late 1930s contri-buted to the Japanese attack on Pearl Harbor.5

5 On the Athenian boycott, see Thucydides, History of the Peloponnesian War, p. 73;Ellings (1985), pp. 17±18, and HSE (1990a), pp. 4±5. On the uses of economic coercion

6

Introduction

The end of the cold war has sparked a renaissance in the use ofeconomic statecraft. The United States has been the most prominentand proli®c actor to employ economic coercion. The National Associ-ation of Manufacturers asserts that between 1992 and 1996 the UnitedStates imposed or threatened economic sanctions sixty times againstthirty-®ve different countries, affecting 42 percent of the world'spopulation.6 These sanctions are estimated to exact an annual cost ofclose to $20 billion in lost exports.7 Richard Haass, writing in ForeignAffairs, goes further, observing: ``What is noteworthy . . . is not just thefrequency with which sanctions are used but their centrality;economic sanctions are increasingly at the core of US foreign policy.''8

The United States is the most noticeable actor employing economicstatecraft; it is hardly the only one. The United Nations SecurityCouncil implemented sanctions seven times in 1994 alone, as opposedto mandating sanctions only twice in its ®rst forty-®ve years.9 TheRussian Federation has employed economic coercion as a way ofextracting political concessions from the Baltic republics and theCommonwealth of Independent States. Lesser powers such as Greece,Turkey, and Nigeria have used economic coercion as an element oftheir regional foreign policies. Even non-state actors are employingthis tool. In December 1996, De Beers threatened to boycott allpurchases of Russian diamonds unless the government acquiesced togranting the South African company monopoly control over its rawdiamond exports.10

There is every reason to believe that the prominence of economiccoercion will increase in the future. Over the course of the pastcentury, major powers have been increasingly reluctant to use orthreaten force, while at the same time demonstrating a growingeagerness to employ economic coercion.11 Bosnia, Chechnya, andSomalia have highlighted the costs of military intervention for the

between 1400 and 1800, see Ellings (1985), pp. 18±21 and Irwin (1991). On the Anglo-American economic warfare, see Renwick (1981), chapter 1. For Wilson's belief in thepower of economic sanctions, see Daoudi and Dajani (1983), p. 26; for cases of League ofNations sanctions, see Doxey (1980), chapter 4. Hirschman (1945) provides the bestaccount of Nazi economic statecraft in the 1930s. The US embargo of Japan is discussedin HSE (1990b), pp. 53±61.

6 Schlesinger (1997), p. 8.7 Hufbauer, Elliott, Cyrus, and Winston (1997).8 Haass (1997), p. 74. 9 Pape (1997).

10 OMRI Daily Digest, ``De Beers issues ultimatum to Russia,'' December 19, 1996.11 Pollins (1994).

Introduction

7

great powers. Unless the use of force is quick and successful, militar-ized disputes sap a nation's resources and create a domestic politicalbacklash against the sender government. As public resistance tomilitary interventions increases, and as foreign aid budgets areslashed, policy-makers are turning more and more to economiccoercion as an attractive substitute to advance the national interest.12

With the increased popularity of this policy tool comes the need fora better understanding of how it works. Analyzing foreign policy islike honing a knife. A sharper knife makes a cleaner cut; a well-understood policy option makes for well-executed policies. However,as the next section will show, the existing literature is of limited use topolicy-makers. Most of the scholarly and policy discussions ofeconomic sanctions consist of debates about high-pro®le cases. Policy-makers have a disturbing tendency to ignore explicit theory but to useanalogies to celebrated cases as a poor substitute. A clear model ofcoercion can provide statesmen with a strategic knowledge that,combined with their knowledge of the speci®c dispute, allows them tobridge the gap between theory and policy.13

The second reason for studying economic coercion is to address therelationship between power and interdependence. Power is the cur-rency of world politics. International relations theorists have alwaysappreciated the power of the sword, but disagree about the import-ance, utility, and de®nition of economic power. Modern realismdeveloped in reaction to the Wilsonian faith in economic power toregulate international politics. Not surprisingly, realists tend to deni-grate the utility of economic statecraft. Neoliberal institutionalismdeveloped in reaction to the realist paradigm; neoliberals believe thateconomic interdependence can affect the behavior of nation-states forthe better.14

This debate is not trivial. If economic sanctions are a potent tool ofdiplomacy, then world politics can be much less violent than it was inthe past.15 Neoliberals argue that increased interdependence in themodern world will cause states to act in a more cooperative fashion,because it increases the costs of defection. The prisoner's dilemmashows the importance of sanctions to neoliberals. In a world full of

12 Rogers (1996). 13 George (1993).14 Keohane and Nye (1978).15 This is not meant to imply that economic coercion has no human costs. The UNsanctions imposed against Iraq have had a serious humanitarian impact on that nation'scitizenry. See Lopez and Cortright (1997) and Buck, Gallant, and Nossal (1998).

8

Introduction

prisoner's dilemmas, states will go it alone unless they expect to bepunished for defecting. Increased levels of economic interdependence,it is argued, make punitive but peaceful strategies possible in anumber of different arenas. Joseph Nye notes, ``Interdependence doesnot mean harmony. Rather, it often means unevenly balanced mutualdependence. Just as the less enamored of the two lovers may manip-ulate the other, the less vulnerable of two states may use subtle threatsto their relationship as a source of power.''16 For cooperation to be astable outcome, countries must believe that it is best to avoid beingthe target of sanctions.17 Robert Axelrod and Robert Keohane note:``When sanctioning problems are severe, cooperation is in danger ofcollapse . . . To explain the incidence and severity of sanctioningproblems, we need to focus on the conditions that determine whetherdefection can be prevented through decentralized retaliation.''18

Axelrod and Keohane use the term ``sanctions'' to mean a variety ofpunitive measures, but economic coercion would certainly be aprominent example.

Neoliberals assume that potent economic sanctions provide anincentive for cooperation. That assumption cannot go unexamined. Ifneoliberals are correct, then it is possible for the power of the sword tobe trumped by the power of gold. If they are not correct, then statesmay blunder into war because their faith in economic statecraft ismisplaced, and the description of the world as a manageable prison-er's dilemma is inaccurate. The better we understand the dynamics ofeconomic coercion, the better we can evaluate the effect of inter-dependence on international interactions.

The literature

In 1945, Albert Hirschman argued in National Power and the Structureof Foreign Trade that great powers could use economic coercion toextract concessions from weaker states. Hirschman's analysis cappedtwo decades of writings that thought of economic coercion as a potentdiplomatic tool.19 That belief faded quickly with the onset of the cold

16 Nye (1990), p. 158. See also Crawford (1994).17 See Oye (1986), Axelrod (1984), Rosecrance (1986), and Buzan (1984).18 Axelrod and Keohane (1986), p. 236.19 Hirschman (1945); Daoudi and Dajani (1983). See Wagner (1988) for a sophisticatedevaluation of Hirschman's approach.

Introduction

9

war. Since then, pundits and policy-makers have disparaged the useof sanctions in foreign policy:20

George Kennan: ``There have been suggestions that we should with-hold m.f.n. treatment, and indeed discourage trade itself, as a meansof extorting political concessions generally . . . This idea seems to meto be quite unsound; it is in any case impracticable.''

Richard Nixon: ``Some people think of economic leverage as thepunitive use of economic sanctions, with highly publicized condi-tions set for their removal. This is highly ineffective, and sometimescounterproductive.''

George Shultz: ``As a general proposition, I think the use of tradesanctions as an instrument of diplomacy is a bad idea . . . Our using ithere, there and elsewhere to try to affect some other country'sbehavior . . . basically has not worked.''

Milton Friedman: ``All in all, economic sanctions are not an effectiveweapon of political warfare.''

Time: ``Economic sanctions have rarely been successful.''

US News and World Report: ``The problem with sanctions is that, moreoften than not, they fail to achieve results.''

Far Eastern Economic Review: ``Of the many arguments againsteconomic sanctions, we have always found the most persuasive isthe simplest: they don't work.''21

This disdain mirrors the scholarly community's consensus aboutsanctions. David Baldwin, who provides the most authoritativesurvey of prior work, observes, ``The two most salient characteristicsof the literature on economic statecraft are scarcity and the nearlyuniversal tendency to denigrate the utility of such tools of foreignpolicy.''22 A ®rst cut of this literature would seem to con®rm thisassessment. Consider the following statements:

Johan Galtung: ``In this article the conclusion about the probableeffectiveness of economic sanctions is, generally, negative.''

20 Even Hollywood is derisive; in the 1997 movie Air Force One, Harrison Ford, playingthe President, denounces a policy of applying economic sanctions to terrorist states as``cowardly.''21 Nixon, Shultz, Friedman, and Time quotations from Daoudi and Dajani (1983), pp. 47,184±7; Kennan (1977), p. 220; US News and World Report, ``Sanctions: the pluses andminuses,'' October 31, 1994, p. 58; Far Eastern Economic Review, ``Sanctioning Burma,''May 8, 1997, p. 5.22 Baldwin (1985), p. 51.

10

Introduction

Peter Wallensteen: ``[T]he general picture of the sanctions is that theyare highly unsuccessful in bringing about the compliance desired.''

Henry Bienen and Robert Gilpin: ``With very few exceptions andunder highly unusual sets of circumstances, economic sanctions havehistorically proven to be an ineffective means to achieve foreignpolicy objectives.''

Margaret Doxey: ``The record of international sanctions of a non-military kind, even when applied within an organizational frame-work, suggests that on their own they will not succeed in drasticallyaltering the foreign or domestic policy of the target.''

Makio Miyagawa: ``Notwithstanding such serious impacts upon thetarget countries, economic sanctions have only rarely achieved thedeclared goals.''

Abram Chayes and Antonia Handler Chayes: ``When economicsanctions are used, they tend to be leaky. Results are slow and notparticularly conducive to changing behavior.''

Robert Pape: ``[E]conomic sanctions have little independent useful-ness for [the] pursuit of noneconomic goals.''23

These are strong statements from a profession accustomed to hedging.A second cut at the literature reveals two distinct strains of analysis

of the sanctions issue. One set of explanations, the domestic politicsapproach, focuses on the politics within the sender and target coun-tries. The decision to initiate sanctions is caused by the domesticpressure within the sender country. The outcome of a sanctions effortwill most likely be failure because of the domestic politics within thetarget country. The second set of arguments, the signaling approach,focuses on systemic variables to explain why economic coercion isoften imposed but rarely pro®table.

According to the domestic politics approach, if the target country'sbehavior violates international norms, citizens in sender countries willfeel compelled to ``do something.'' Even if the foreign policy leaderagrees with public opinion, the costs of effective military interventionmay be too high. On the other hand, the domestic political costs ofdoing nothing are substantial, because it creates the image of a weakleader. The lack of options leaves the sender regime hamstrung.

23 Galtung (1967), p. 409; Wallensteen (1968), pp. 249±50; Bienen and Gilpin (1980),p. 89; Doxey (1987), p. 92; Miyagawa (1992), p. 206; Chayes and Chayes (1995), p. 2;Pape (1997), p. 93. For other pessimistic observations, see Knorr (1975), pp. 205±6; vonAmerongen (1980), p. 165; Blessing (1981), p. 533; Willett and Jalalighajar (1983), p. 718;Lindsay (1986), p. 154; Hendrickson (1994), pp. 22±3; Barber (1995), p. 29.

Introduction

11

Senders will turn to economic statecraft as an imperfect substitute forforceful action. Economic coercion can de¯ect domestic pressure andregister the sender regime's disapproval of the target's actionswithout going to war. The sanctions themselves might be ineffective,but their implementation allows the foreign policy leader to avoidaccusations of do-nothing leadership.

Many authors have proposed all or part of this rationale. JamesBarber observes: ``The purpose of sanctions here is to demonstrate awillingness and capacity to act. Negatively, the purpose may simplybe to anticipate or de¯ect criticism.''24 M. S. Daoudi and M. S. Dajaniconcur: ``The imposition of sanctions absorbs the initial public reactionthat something needs to be done.''25 Ivan Eland concludes:

Bluntly stated, most of the times a nation imposes sanctions onanother country, it has few policy options. The target nation usuallyhas committed an unacceptable act and intense domestic pressure,particularly in democratic states, to ``do something'' can persuadethe government in the sanctioning nation to respond by imposingsanctions to meet goals other than target compliance.''26

The literature is rife with assertions like these.27 From a foreign policyperspective, rational calculation plays a limited role; from a domesticpolitics perspective, economic statecraft serves as a steam valve torelieve governments from the pressure of their populace. In thisexplanation, sanctions are symbols; their effectiveness is of secondaryconcern.

Domestic-level explanations provide three factors working againstthe utility of economic coercion. First, given the causes of sanctionsimposition, their implementation will be erratic and haphazard. Thus,the target country may not be inconvenienced at all. Second, even ifthe sanctions are potent, target governments can use the specter ofinternational coercion to create a ``rally-round-the-¯ag'' effect. Dom-estic groups line up behind the government in reaction to an externalthreat. To do otherwise would smack of disloyalty. There are psycho-logical factors which reinforce this effect. Johan Galtung's study ofRhodesia noted that the mutual sacri®ces created by the sanctions led

24 Barber (1979), p. 380.25 Daoudi and Dajani (1983), p. 161.26 Eland (1995), p. 29.27 See Renwick (1981), p. 85; Schreiber (1973), p. 413; Hoffman (1967), p. 154; Daoudiand Dajani (1983), appendix II; Haass (1998).

12

Introduction

to an esprit de corps among the citizenry and a closer identi®cationwith the government.28

The third reason is that target governments may, for domesticreasons, prefer to be sanctioned. In the long run, sanctions hurt thetrade-oriented sectors of the economy by depriving them of income.At the same time, an embargo strengthens import-substitution sectorsby giving them rent-seeking opportunities. Since export sectors willprefer the target government to acquiesce, a lengthy sanctions disputecan politically weaken the foreign policy leader's domestic opponents.This is particularly true if the target regime is authoritarian. Sanctionspermit target regimes to strengthen state control over the economy,and readjust the impact of sanctions policies away from its mostpowerful supporters. For example, when the UK-led coalitionimposed sanctions against Rhodesia, household incomes for blackfamilies fell, while white incomes rose. Serbian leader SlobodanMilosevic used the United Nations embargo to reward crony enter-prises with scarce goods, while punishing his political rivals. UnitedNations aid to Iraq has freed up funds for Saddam Hussein to rewardhis inner elite.29

For economic coercion to work, target elites must suffer as much astarget populations. Case studies of Uganda have shown that thesanctions to remove Idi Amin became more effective when GreatBritain halted the export of luxury goods. This hurt the Ugandanarmy elite, which relied on the ``whiskey runs'' for creature comforts.Accounts of the sanctions against Haiti after 1990 revealed that theHaitian military regime was willing to negotiate only after the Clintonadministration prevented the Haitian armed forces from acquiring oilor weapons on the global market.30

The half-hearted motivations of sender governments, combinedwith backlash effects within the target country, make it extremelydif®cult for economic sanctions to generate concessions. A domesticpolitics approach produces several hypotheses, as seen in Table 1.2.Sanctions are more likely to be initiated when the sender is a

28 Galtung (1967).29 On Rhodesia, see Losman (1979), pp. 112±13, and Rowe (1993). On Serbia, seeWoodward (1995), p. 148 and Licht (1995), p. 158. On Iraq, see Wright and Daniszewski,``Hussein may be main bene®ciary of UN aid effort,'' p. A1, and more generally, Reuther(1995), pp. 125±7.30 On Uganda, see Miller (1980), p. 124, and Ullman (1978), pp. 532±3; on Haiti, seeWerleigh (1995), p. 168. More generally, see Morgan and Schwebach (1996).

Introduction

13

democracy. Public opinion to do something should resonate morewith foreign policy leaders facing electoral pressures. It is also arguedthat sanctions will be used primarily when the sender cannot usemore persuasive means of statecraft. Therefore, states will useeconomic coercion when the costs of military intervention are toogreat. If the target is physically distant, power projection becomes amore dif®cult enterprise, and sanctions are therefore more likely.31

31 This hypothesis does not test domestic-level variables, but still comes from adomestic-level approach. An implicit assumption of this level of analysis is thatinternational factors constrain the sender regime from acting more forcefully. Therefore,this hypothesis must be true for domestic factors to have an appreciable effect.

14

Introduction

Table 1.2. Existing approaches to economic coercion

Domestic politics Signaling

Causes of coercionattempt

1. Domestic pressure onthe sender regime

2. Lack of palatablealternatives

1. Desire to signalfuture actions

Causes of coercionoutcome

1. Ability of targetregime to use sanctionsto its own politicaladvantage

1. Cost of the signal tothe sender

2. Implicit threats ofpower projection ormilitary force

3. Ability of the sendercountry to attractmultilateralcooperation

Coercion is more likelyto be attempted if:

1. The potential senderis a democracy

2. The target isgeographically distantfrom the sender

1. No predicted pattern

Coercion is more likelyto generate concessionsif:

1. The target regime isdomestically unstable

2. Sanctions hurt thetarget elites as much asthe general population

1. The sanctions arecostly to the sender

2. Military force is alsothreatened or used

3. The sender attractsinternationalcooperation

Two predictions can also be made about the likelihood of sanctionssuccess. First, if the target government is domestically unstable, it maylack the means to convert a sanctions dispute into political support.Such a government would be more likely to acquiesce so as to hold onto power, or be removed in favor of those who prefer accommodation.Second, if target elites are made to suffer as much as target popula-tions, there is no opportunity for rent-seeking, which puts elitepressure on the target government to concede.

The signaling approach to economic statecraft has little to say aboutthe initiation of sanctions, but pays more attention to variablesaffecting the outcome. Two factors are frequently cited. The ®rst is thebackground assumption that without a high degree of internationalcooperation, sanctions are useless. International trade theory suggeststhat for a homogeneous good with a high substitution elasticity, only asender coalition responsible for more than half the supply of thatgood can in¯uence the terms of trade.32 Few individual sendercountries have this capability, and when they do it is usually ephem-eral. Therefore, for sanctions to have any in¯uence, internationalcooperation is necessary. This explains the overwhelming focus in theliterature on multilateral cases of economic coercion. Obtaining inter-national cooperation is exceedingly dif®cult, however.33 As morecountries join in the coercion attempt, the sanctions coalition getsmore unwieldy. There is a greater incentive for individual countries tofree ride, permit illicit trade, and pocket increased pro®ts.34 Becauseof the dif®culties in sustaining multilateral cooperation, the signalingapproach is skeptical about the prospect of economic coercion suc-ceeding on its own merits.

While sanctions rarely generate concessions, they can function aseffective signals. This argument rests on the assumption that statesconduct foreign policy in a world of imperfect information. If statesare uninformed about other states' preferences, there is always anincentive to bluff in international crises. For example, if the United

32 Gardner and Kimbrough (1990).33 See Bayard, Pelzman, and Perez-Lopez (1980), Doxey (1980, 1987), Martin (1992),Mastanduno (1992), Mans®eld (1995), Kaempfer and Lowenberg (1997), and Drezner(1998) for more on cooperation and economic sanctions.34 Even if states nominally comply with multilateral sanctions, individual ®rms may betempted. This was certainly the case with Yugoslavia's neighbors in the early 1990s. SeeRaymond Bonner, ``How sanctions bit Serbia's neighbors,'' New York Times, November19, 1995, section 4, p. 3.

Introduction

15

States threatens to use force against a target country and the targetacquiesces, it has won without having to carry out its threat. Thepossibility of a costless victory creates an incentive to make evenempty threats. Unless the target country knows the extent of USwillingness to use force, it will have dif®culty distinguishing betweena credible threat and cheap talk.

Because of imperfect information, states frequently engage in sig-naling techniques to demonstrate credibility. Some acts, such as themobilization of troops, can signal that rhetoric will be translated intoaction if the sender's demands are not met. Economic sanctions can bethought of as another type of signal. The key to a successful signal isto take an action that will separate credible threats from cheap talk.For this reason, a costly signal is better than a cheap signal. If a signalis costly, a bluf®ng sender is less likely to use it because of its price.35

According to this argument, economic sanctions are ineffective ascoercive tools, but may be useful as signals. Their value as a signalcomes not from the damage in¯icted on the target, but the cost to thesender. David Baldwin notes in Economic Statecraft: ``Other thingsbeing equal, it is always desirable to minimize costs; but other thingsare not equal. The selection of a costly method of conveying a signalmay add credibility to the signal. Thus, a statesman interested indemonstrating resolve may want to avoid the less expensive means ofcommunication.''36 In a world of cloudy signals, policies that provecostly to the sender can be an excellent means of conveyinginformation.

Lisa Martin makes a similar argument about the relationshipbetween the sender's costs and acquiring multilateral cooperation forsanctions. She observes that potential allies in the sanctioning effortneed to be convinced of the sender's commitment. Costly sanctions bya great power can convince other states to join in the sanctioningeffort; the high costs act as a signal of the great power's seriousness ofintent.37 Thus, a costly signal of sanctions helps to send a crediblesignal to other possible senders as well as the target.

Note the implicit argument that runs through this entire line ofreasoning. According to this logic, unilateral economic sanctionscannot work on their own. They are only effective if they act as asignal that stronger measures, like multilateral embargoes or military

35 Schelling (1960), Fearon (1994).36 Baldwin (1985), p. 372. 37 Martin (1992), pp. 36±8.

16

Introduction

intervention, will be taken in the future. The causal argument in thisschool of thought is that what appears to be a sanctions success isactually the product of an implicit military threat. Sanctions, therefore,are not a true cause of concessions, but merely an observable signal ofmilitary power.38

This logic is consistent with empirical claims that the few successesascribed to economic coercion are really examples of successfulmilitary threats.39 Recent examples also provide support. US sanctionsagainst Haiti from 1990 to 1994 were successful in removing themilitary junta in power and reinstalling Jean-Bertrand Aristide aspresident. Through the fog of history, it would be easy for futurescholars to argue that the sanctions caused the outcome. That wouldobviously overlook the crucial role of the US military threat. TheHaitian leadership acquiesced when they were told by Americannegotiators that the US 82nd Airborne Division was on its way.Sanctions may have assisted in the return of Aristide, but the chiefcause was the threat of force.

Table 1.2 shows the empirical tests that can be derived from asignaling explanations. There are clear hypotheses about the like-lihood of sanctions success. First, if the sender incurs signi®cant costsin its sanctions attempt, the credible signal is more likely to produceconcessions. If the signal fails to work, then economic coercion willonly work if it is associated with companion policies that genuinelyproduce concessions. Multilateral cooperation in the sanctions effort isexpected to generate concessions. Also, if there are threats of militaryor quasi-military intervention in the dispute, the target is more likelyto acquiesce for obvious reasons, even though economic coercion isnot the cause.

Reviewing the literature, one can see a con¯uence of domestic andinternational factors leading to the same conclusion: economic state-craft rarely works. Sanctions are initiated because other options arenot feasible, and the sender regime wishes to placate domesticpressures to take action. They have little chance of success. Domesticpolitics within the target country will lead to de®ance. Unless sanc-tions attract signi®cant multilateral cooperation, and hurt target elitesas much as the population, they will not force acquiescence. Sincethese conditions are dif®cult to achieve, economic pressure will rarely

38 Lenway (1988), and Morgan and Schwebach (1997) also make this observation.39 Knorr (1975), Schreiber (1973), and Pape (1997).

Introduction

17

be the causal mechanism for target concessions. More likely, costlysanctions act as a signal of stronger measures and convince the targetto back down.

Flaws in the literature

And that, it would appear, is that. It would be comforting to think thatpolitical scientists have successfully described at least one corner ofthe foreign policy arena. It would also be wrong.

Economic sanctions are more effective than the literature claims.Gary Hufbauer, Jeffrey Schott, and Kimberly Ann Elliott have createda database of 116 sanctions cases.40 They code the success of theepisode, as traditionally de®ned, into one of four categories, rangingfrom outright failure to complete success. If the most stringentde®nition of success is used, more than 35 percent of the cases qualifyas a success. Less than 30 percent of the cases listed fall into thecategory of complete failure. In many of these successes, there was noobserved threat of military force, no multilateral cooperation, and norwere the sanctions particularly costly to the sender. This evidencehardly suggests that sanctions always work, but it calls into doubt thehypothesis that they always fail.

Why is there such a discrepancy? The ®rst problem is one ofde®nition. The question phrased in the literature is ``do sanctionswork?'' This is too simplistic. M. S. Daoudi and M. S. Dajani comment:``Most studies have assumed that the objectives of economic sanctionswere to return to the status quo that prevailed prior to the act ofaggression which brought the sanctions about. In reality, the aims ofsanctions have been consistently less ambitious.''41 Scholars comparethe sanctions outcome with the status quo ante (i.e. before the actionthat prompted sanctions) to determine the success of a coercion event.This is the wrong counterfactual; the hypothetical alternative is theoutcome if coercion was not attempted in the ®rst place. If the targetedcountry does not change its policies at all, then the event should bejudged a failure. If there is some compromise, however, and the valueof the concession outweighs the sender's costs of coercion, then theepisode counts as a partial success.

The degree of success also depends upon the type of demand.

40 HSE (1990a, 1990b). Pape (1997) argues that many of these cases are miscoded.41 Daoudi and Dajani (1983), p. 2.

18

Introduction

Baldwin notes: ``A moderate degree of success in accomplishing adif®cult task may seem more impressive than a high degree of successin accomplishing an easy task. In assessing statecraft, as in judgingdiving contests, scores should be adjusted for the level of dif®culty.''42

These nuances are overlooked in the simple dichotomy of success/failure made in the literature.

The second problem is the tenuous link between internationalrelations theory and the sanctions literature. Most of the recentcontributions on economic statecraft consist of well-crafted theoriesthat lack empirical support,43 or well-crafted case studies thatproduce generalizations of dubious quality.44 The literature oftenoverlooks theoretical developments that blunt the utility of theircausal mechanisms. Consider the effect of domestic politics on boththe sender and the target country. It is theoretically unclear whetherthe sender is more likely to succeed if there is domestic support forsanctions or if there is some political opposition. On the one hand, thetwo-level games approach argues that unanimous domestic supportenhances the sender's bargaining position because it reduces thelikelihood of the sender reversing its strategy.45 On the other hand, anincomplete information approach would make the opposite claim:domestic audience costs send an effective signal of resolve to thetarget country. Now consider the effect of domestic politics within thetarget country. The domestic politics approach argues that the morevulnerable target elites are to sanctions, the more likely that they willacquiesce. A rational choice perspective would suggest an alternativeoutcome, however. If sanctions narrow the ruling coalition within thetarget country, it could force the decision-maker into a more hardlinebargaining position. A weakened leader who cannot afford to alienatea narrow coalition of hardliners will stand ®rm. A leader with abroader base of support, or one insulated from public opinion, has theluxury to concede without fearing the collapse of the target regime.

These contradictory effects within the target and sender countriessuggest that the sanctions literature exaggerates the absolute effect ofdomestic politics on economic coercion. It is possible that both thetwo-level games approach and the audience cost approach are correct,but that in most situations, the effects cancel each other out. Similarly,

42 Baldwin (1985), p. 372.43 See Tsebelis (1990), Eaton and Engers (1992), Smith (1996).44 Cortright and Lopez (1995), Klotz (1996).45 Moravscik (1993), pp. 29±30.

Introduction

19

the opposing effects of sanctions within the target country's rulingcoalition could lead to the overall insigni®cance of domestic politics.Empirically, this would mean that although isolated cases wouldshow the importance of domestic politics, the aggregate effect wouldbe minor.

Finally, the sanctions literature is also guilty of numerous methodo-logical sins.46 First, almost all of the arguments use an inductiveapproach; theories about economic statecraft are developed only afteran examination of case studies. Under Arend Lijphart's typology,most of these works are interpretive case studies.47 The concern is notwith theory, but with explaining the speci®c event. As such, few of thederived propositions have the necessary fecundity to be useful inother issue areas. The inductive approach causes researchers to ignoreimportant variables or questions raised by broader approaches tointernational relations. Case selection exacerbates these problems.Most of these writings focus on the more celebrated cases of sanctions.Open a book on economic sanctions and most of its pages will bedevoted to the following cases: the League of Nations sanctions onItaly; CoCom's strategic embargo of technology exports to the com-munist bloc; and UN sanctions against Rhodesia, South Africa, Iraq,and/or Serbia. Several theories of economic coercion have beendeveloped almost exclusively from these cases.48

Certainly these sanctions cases are well known, but they do notnecessarily represent the universe of observations. Their very celebritysuggests they are atypical; they are important because they stand outin some unusual way. In most of the cases, the primary sender hadmultilateral assistance and backing from an international organiza-tion. The demands made of the target country were non-negotiable.The primary sender and target were adversaries. The sanctions policyusually failed. By choosing cases that take on extreme values of boththe dependent variable and several independent variables, the litera-ture commits two errors. First, there is a tendency to underestimatethe main causal effects on the universe of events. Second, these

46 In many ways the problems are similar to the extensive literature on deterrencetheory. The criticisms presented here also have a parallel in that literature. See Achenand Snidal (1989).47 Lijphart (1971).48 See Adler-Karlsson (1968), Losman (1979), Renwick (1981), Doxey (1987), Leyton-Brown (1987), Mastanduno (1992), Cortright and Lopez (1995), Haass (1998).

20

Introduction